lv_i18n - Internationalization for LittlevGL

Lightweight gettext replacement tools for C. Add multi-language support to your embedded projects with ease.

Quick overview

1. Mark up the text in your C files as _("some text") (singular) and _p("%d item", item_cnt) (plural)
2. Create template yml files for the translations you are interested in
3. Run extract to fill the yml files with the texts in _() and _p()
4. Add the translations into the yml files
5. Run compile to convert the yml files to a C and H file. They will contain the translations and all the background functions you need.
6. Be sure your fonts contain the required characters. See the docs here for more details.

Install/run the script

node.js required.

Global install of the last version, execute as "lv_i18n"

npm i lv_i18n -g

Alternatives:

Install from github's repo, master branch

npm i littlevgl/lv_i18n -g

If you wish local install for your project, do in project root:

npm init private
npm i lv_i18n -s
# now available at ./node_modules/.bin/lv_i18n

Then commit package.json & put /node_modules into .gitignore. Next time use just npm i to install.

run via npx

node.js has built-in npx utility to execute packages "without install":

# run from github master
npx github:littlevgl/lv_i18n -h
# run from npm registry
npx lv_i18n -h

Mark up the text in your code

#include "lv_i18n/lv_i18n.h"  /*Assuming you have the translations here. (See below)*/

/* Load translations & default locale (usually, done once) */
lv_i18n_init(lv_i18n_language_pack);

/* Set active locale (can be switched anytime) */
lv_i18n_set_locale("ru-RU");

/* The translation of "title1" will be returned according to the selected locale.
 * ("title1" is only a unique ID of the text.) Example:
 * en-GB: "Main menu"
 * ru_RU: "Главное меню"
 */
gui_set_text(label, _("title1"));

/* According to `user_cnt` different text can be returned
 * en-GB `user_cnt == 1` : "One user is logged in"
 *        `user_cnt == 6` : "%d users are logged in"  
 */
char buf[64];
sprintf(buf, _p("user_logged_in", user_cnt)), user_cnt);  /*E.g. buf == "7 users are logged in"*/
gui_set_text(label, buf);

_ and _p are normal functions. They just have this short name to enable fast typing of texts.

Rules of getting the translation:

• If the translation is not available on the selected locale then the default language will be used instead
• If the translation is not available on the default locale the text ID ("title1" in the example) will be returned

Create template yml files for the translations

For each translation, you need to create a yml file with "language code" name. For example:

• en-GB.yml
• ru-RU.yml

Here is a list of the language and locale codes.

Add the locale-name: ~ line to the yml files. Replace "locale-name" with the actual language code.
E.g.: en-GB: ~ or simply en: ~

Technically you can have one yml file where you list all language codes you need but its more modular to separate them.

Run extract to fill the yml files

Run extract like this (assuming your source files are in the src folder and the yml files in the translations folder):

lv_i18n extract -s 'src/**/*.+(c|cpp|h|hpp)' -t 'translations/*.yml'

It will fill the yml files the texts marked with _ and _p. For example:

en-GB:
  title1: ~
  user_logged_in:
    one: ~
    other: ~

Add the translations into the yml files

The naming conventions in the yml files follow the rules of CLDR so most of the translation offices will know them.

Example:

'en-GB':
  title1: Main menu
  user_logged_in:
    one: One user is logged in
    other: '%d users are logged in'

If translators want to know where a message comes from, then use lv_i18n extract --dump-sourceref sr.json ... to generate the file sr.json containing file names and line number of each message.

Run compile to convert the yml files to a C and H file

Once you have the translations in the yml files you only need to run the compile to generate a C and H files from the yml files. No other library will be required to get the translation with _() and _p.

Running compile:

lv_i18n compile -t 'translations/*.yml' -o 'src/lv_i18n'

The default locale is en-GB but you change it with -l 'language-code'.

You can use --optimize to generate optimized C code. Without this, finding the corresponding translation by lv_i18n_get_text() is done by searching through all keys until the right one is found. This can eat up a lot of CPU espcially if the list is long (aka O(n)). Using --optimize changes this behaviour by using an integer index into the list of strings resulting in an immediate return of the right string (aka O(1)). As the index is computed at compile time, you need a compiler, which is able to evaluate strcmp() with constants at compile time. All modern compilers, like gcc and clang are able to do this. If you want to check, whether your compiler is able to handle this optimization, you can use the following code to check this:

int main()
{
  return strcmp("a", "a");
}

If this compiles without needing #include <string.h> and nm -u a.out does not output strcmp as being undefined, then the compiler optimizes the code and is able to handle --optimize.

Follow modifications in the source code

To change a text id in the yml files use:

lv_i18n rename -t src/i18n/*.yml --from 'Hillo wold' --to 'Hello world!'

C API

int lv_i18n_init(const lv_i18n_language_pack_t * langs)

Attach generated translations to be used by lv_i18n_get_text().

• return - 0 on success, -1 on fail.

---

int lv_i18n_set_locale(const char * l_name)

Set locale to be used by lv_i18n_get_text().

• l_name - locale name (en-GB, ru-RU).
• returns - 0 on success, -1 if locale not found.

---

const char * lv_i18n_get_text(const char * msg_id)

Mapped to _(...) or _t(...) via #define

Get translated text. If not translated, return fallback (try default locale first, then input param if default not exists)

• msg_id - The ID of a text to translate (e.g. "title1")
• return - pointer to the traslation

---

char* lv_i18n_get_text_plural(char* msg_id, int32_t plural)

Mapped to _p(...) or _tp(...) via #define

Get the plural form of translated text. Use current locale to select plural algorithm. If not translated, fallback to default locale first, then to input param.

• msg_id - The ID of a text to translate (e.g. "title1")
• plural - number of items to decide which plural for to use
• return - pointer to the traslation

References:

To understand i18n principles better, you may find useful links below:

• gettext
• Rails Internationalization (I18n) API